subchannel13 / ikon-library Goto Github PK

Add or change methods in IkadnReader to make it possible for 
IIkadnObjectFactory to get indentation level of an object it is about to parse.

Ikadn.Ikon.Types.IkonArray currently has constructor that accepts 
IList<IkadnBaseObject>. Examine whether IList is necessary or it can be 
abstracted to IEnumerable.

References are currently only feature in IKON that reserve a certain character 
for a certain type of data and they are not needed in every IKON based object 
notation.

Consider demoting them as IKSTON exclusive feature.

Split current class for numeric data to multiple classes, each for specific 
type of numbers. Initial proposal is to make three classes: one containing 
long, other with double and third with decimal.

IkonArray.To<T>() method should support conversion to IList<U> where U is valid 
conversion type for underlying list element. For instance 
IkonArray.To<IList<string>>() should succeed if all elements of internal list 
are IkonText.

IKSTON should expose the of more direct way of converting program data to IKON 
values and other way around. Current methods require a lot of type checking and 
casting in the user's code.

Reduce required .Net framework version to v3.5

Introduce a data type/syntax to IKON that would handle multiline text blocks 
better than current textual type.

Parser's ParseAll() method currently returns queue. Common use case is to 
dequeue only values of certain type.

The method should return a type that supports querying for a value of specified 
type.

Representation of special numeric values (not a number, positive and negative 
infinity) should not be language specific. Current version uses representation 
specific to System.Double from .Net framework.

Consider this code snippet:

var array = new IkonArray();
array.Add();

Add method without parameters should not be valid. Try solve it in a way that 
would make compiler catch it as an error.

Check whether negative number format ignores regional settings.

{First}

{Second 
   data []
}

{Second 
   data {First}
}

Error reporting could be enhanced a little bit. Currently (in both IKON and 
IKSTON) is just exception being thrown from somewhere within the parser.

Error report should ideally inform about line and column of error and what was 
expected input.